save notebook

Author: glemaitre

content/python_files/feature_engineering.py

@@ -820,19 +820,24 @@ def splitter(X, y, index_generator):
 ts_cv_2 = TimeSeriesSplit(
     n_splits=2, test_size=test_size, max_train_size=max_train_size, gap=24
 )
-randomized_search_hgbr = hgbr_predictions.skb.get_randomized_search(
-    cv=ts_cv_2,
-    scoring="r2",
-    n_iter=100,
-    fitted=True,
-    verbose=1,
-    n_jobs=-1,
-)
+# randomized_search_hgbr = hgbr_predictions.skb.get_randomized_search(
+#     cv=ts_cv_2,
+#     scoring="r2",
+#     n_iter=100,
+#     fitted=True,
+#     verbose=1,
+#     n_jobs=-1,
+# )
+# # %%
+# randomized_search_hgbr.results_.round(3)
+
 # %%
-randomized_search_hgbr.results_.round(3)
+# fig = randomized_search_hgbr.plot_results().update_layout(margin=dict(l=200))
+# write_json(fig, "parallel_coordinates_hgbr.json")
 
 # %%
-randomized_search_hgbr.plot_results().update_layout(margin=dict(l=150))
+fig = read_json("parallel_coordinates_hgbr.json")
+fig.update_layout(margin=dict(l=200))
 
 # %%
 # nested_cv_results = skrub.cross_validate(